Skills reworked

In 5e, if you are proficient in a skill, tool, vehicle or instrument you add a level based proficiency bonus, and if you get expertise from somewhere (there are a few sources, but the best known is rogue), you double this. You get proficiencies from some races, all backgrounds, and a variable number from your class. You occasionally get some from your subclass. You can also get them from feats, such as Skilled and UA skill feats as well. It's incredibly rare to get more after third level, except from feats.

In 3e/3.5e, skills worked rather differently. At every level you had a number of skill points to spend, based on your character class and Int modifier, and if you bought class skills you got +1 per point you spent, non-class skills gave you +1 per 2 points you spent. So a bard would get +1 per point in Speak Other Language and Perform say, while a Monk might get it in Acrobatics, Tumble and Religion and a Wizard in Arcana and History. You could focus down and improve the same skills every level, ending up with a 20th level monk having a tumble skill of +28 or more (easily) or diversify and have more skills with lower bonuses.

I have to say, I think both methods have pros and cons. 

I like that 5e has a more tightly bounded approach. You're going to get to +7 maximum, from your level, +5 normally from your stats, and that's it. I dislike that you pick your skills and for most classes by first levels, rarely by third level with a subclass, that's it. Done. You can never learn anything new, just get better at what you know. Ok, you can take a skills feat, but, honestly, in 5e who is going to? You don't get enough feats to do that after all. 

Contrarywise, I like that 3e gives you flexibility to improve everything, including skills outside your core class proficiencies if you really want to, I like that it makes Int important, but I'm less keen on the skill inflation. The gap between the haves and the have-nots becomes extreme, even at medium levels, where you might have +10 from levels, +6 from stats +4 from a magic item boosting that skill (which is cheap in 3e) and be getting +20 on a roll at 10th level easily. Unless you roll a nat-1, they roll a nat-20 you're just a tier above them, every time.

So, how would I approach this? 

As is so often the case, I have two approaches, and they require some play-testing;

1. This approach veers more towards the 3e idea, although you can see a 5e influence too. You have a level based proficiency bonus, notionally 1/2 level, rounded up, although that could be reduced to something like 1/3 of level, rounding normally (so 1/3 rounds down, 2/3 rounds up). You still get points to spend per level, as in 3e, based on your class and your Int modifier, and you have core class and other skills, as in 3e, but the highest bonus you achieve in any skill is +5 whether that's by spending 5 points for a core skill or 10 for an other skill. 
2. By contrast this approach veers more towards 5e. In 5e, every background gives you proficiency in 2 skills and 2 other things (languages, tools, vehicles, performance styles etc.). At various points, for the sake of familiarity, I'm going to suggest levels 5, 11 and 17, although I think probably more often than that, you get this again. The balance might need to be different - any balance of 3 and 1 or 2 and 2 say, as long as the DM agrees. If they're more regular, say every other level, a different approach of 1 and 1 or sticking with 2 and 2 and the five skill tiers outlined in #1 above might make more sense. This last is my preferred option, giving you up to a +5 bonus, with classes that currently give you expertise getting bonus skill points to spend. You would still get a level-based proficiency bonus, broadly similar to the current 5e system.

Both of these allow your character to grow and improve as they level up, which I like. 

The biggest single issue I have with this, upfront, is that it apparently divorces the skills approach from the combat skills approach. However, I'm not sure that's necessarily true, or rather it doesn't have to be true.

Traditionally, the fighter class does very poorly for skills picks because they get all those weapon proficiencies "free", essentially as hidden picks the system gives them automatically. We could unpack that, and give the fighter, the paladin, the ranger, the wizard, the cleric etc. the same number of "picks" as the skill monkey characters like the rogue and bard. We just temper it, so they're less versatile wizards, clerics and fighters but more skill-monkey like. We've looked at this from the point of view of invocations and feats already, why not have it here too? So we'd generally split weapons into groups, maybe, swords, axes, maces, hammers, great-weapons, shields etc. We would add to that proficiency is schools of magic maybe (for the caster types). And then skills. Everyone gets a fixed number of picks (say 5) + Int modifier + Dex/Str Modifier (their choice). 

As a game, D&D proclaims that Fighters are masters of all weapons. In practise, how often do we see that? You build your S&B fighter, your GWF, your 2WF and off you go. Racking your invocations where you want them for that in 6e would already help your specialise, being able to stack up your weapon skills too, and ignore the rest that you just don't use, putting them out there for other skills that you might, that could be useful. And there's nothing to stop you being a multi-weapon generalist. You could insist that if you have a positive Int modifier that number of points must be spent on skills or magic, and the Dex/Str must be used on Weapon Skills. If the variable Initiative stat system outlined here is introduced, it might make more sense to use that instead of Dex/Str and say "to be used on class attack skill" and allow "Int or spell-casting stat" for skills, tools, and spell-casting abilities.

Either of these approaches requires some thinking about the Jack of all Trades bard ability. One approach is just to write it out of course - because everyone can pick up everything, why have it? But I like JofT and actually I'm more inclined to remove the Skilled feat and make JoaT more widely available as a feat (or maybe an invocation) that just to bards. I think the simplest approach is to say once you've got it, you have a +0 skill bonus, but addd the level part of your proficiency bonus to all your skills. If we assume the 1/2(1/3 in brackets) level bonus, at mid levels you'd get about +5(+3-4) from your level, but miss the +4-5 a seriously skilled character would have, your stats would be about the same, for up to +4, so you'd be about in a range from +4 to +8 rather than +12 compared to the good characters. At 20th levels you'd get +10(+7) and miss the +5 skill. You'd have the same stat, which would probably be +8 for a good one, might still be 0 for a bad one, so you might be +10(+7) to +18(+15) where the trained character would be +23(+20). The trained character might have, if the magic item market is there and the 3e skill-boost magic items have come over, an item to boost that skill by +1-+5 though. There's still a lag, but not as pronounced on the good stat skills, but in your poor stats, you lag appreciably. I don't think this is too bad.

Fudge D&D

Since Fudge D&D is a complete rewrite, this could easily be fitted in. It's not a hard patch to do.